1. Field of the Invention
The present invention pertains to communication systems and methods for communicating information over an Integrated Services Digital Network (ISDN) network. In particular, the present invention pertains to establishing over an ISDN D channel a communication link between a source terminal and a destination terminal via a modified ISDN switch that handles two ISDN B channels as a single composite B channel without requiring the source and destination terminals performing bonding or multilink protocol processes.
2. Discussion of the Background
Conventional facsimile devices communicate over the Public Switch Telephone Network (PSTN) using analog signals that are transmitted over conventional telephone lines. The source terminal (e.g., a facsimile device, computer with scanner and modem facilities, or another device that transmits and/or receives data) converts digital scanned information into a corresponding analog signal so the same may be sent over the PSTN telephone line, via a telephone switch facility, to the destination terminal. The source terminal receives the analog information and converts the analog information back into digital signals which form the basis of an image to be printed, perhaps on facsimile paper.
The Integrated Services Digital Network (ISDN) is emerging as a next generation worldwide public telecommunications network that will replace existing public switch telephone networks and provide a variety of services that are not offered by the PSTN. ISDN will allow the transmission of various types of data between various types of ISDN terminal equipment (TE).
A portion of the ISDN link between a source terminal and a central office, which has a switch facility, is referred to as a "digital pipe". A capacity of the pipe is generally discussed in terms of separate channels. In particular, a "basic access" digital pipe includes two B channels (basic channels) that each support 64 kbps signaling, and a D channel at 16 kbps. While the total bit rate of these three channels is 144 kbps, framing, synchronization and other overhead bits bring the total bit rate of a basic access link to 192 kbps. Furthermore, the B channels serve as separate communication channels such that the maximum data capacity, as view by the user, is 64 kbps per B channel, and 16 kbps for the D channel, not 192 kbps.
Conventional facsimile devices at most send signals at rates not exceeding 64 kbps, because only one of the two B channels are used. Because facsimile data is arranged in a predetermined format, sending data over two separate B channels would be a sizable task because conventional ISDN switches handle the B channels separately, and thus may send data of one of the B channels over a completely different route than that of the other B channel. As a consequence, the different communication paths impose different communication delays on the respective B channels.
Other devices such as video teleconference facilities, assume the processing burden of "bonding", or employing multilink point-to-point (multilink PPP) protocols so as to increase the data rate approaching a 128 kbps digital connection. The bonding approach imposes on the customer premise equipment (CPE) the function of dialing to the ISDN switch and establishing the subsequent calls needed to achieve the desired data rate. Thus, two separate links are established. In particular, by assuming the burden of maintaining two separate communication connections with the ISDN switch, the CPE can give the appearance to a user that a 128 kbps channel is available to the user. However, the bonding approach is cumbersome in that the ISDN switch assumes each of the B channels may be handled independently, and therefore impart different delays into the separate B channels. As a consequence, the CPE must compensate for the delays between the respective B channels, and piece together the received and transmitted information so as to avoid synchronization problems.
Multilink PPP schemes attack the problem from a different approach, although also placing a similar processing and data management burden on the CPE. The multilink PPP attempts to make the ISDN switch oblivious that two B channels are effectively being combined to provide an effective data rate approaching 128 kbps. The multilink protocol involves dividing the user's source data into specific fragments, including overhead information in the respective packets, so that the packets may be sent over all available channels, and later recombined in a contiguous fashion. As with bonding, multilink PPP places a computational and management burden on the CPE, rather than at the ISDN switch.
Conventionally, the function served by the ISDN D channel, is twofold. First, the D channel is used to establish and maintain signaling between the CPE and the ISDN switch (operated by the telephone company). Thus, the D channel carries signaling information such as that required for dialing the telephone number of the destination terminal and making the connection between the source terminal and the destination terminal. A more complete description of the D channel as employed in narrowband and broadband ISDN, as well as ISDN terminal equipment, protocols, data rates, etc. is provided in the literature, for example in Stallings, W., "Data and Computer Communications", 5.sup.th Edition, Prentice Hall, 1997, pp 740-769 (hereinafter "Stallings") the contents of this book being incorporated herein by reference.
FIG. 1 is a block diagram of a conventional ISDN system having a source facsimile 10 at a source facility 1 that communicates via an ISDN switch 22 to a destination facsimile 16 (or other type of destination terminal, such as a computer, ISDN equipped photocopier, etc.) in a destination facility 2. The source facsimile 10 communicates via a terminal adapter 10A, shown as an internal device, although a separate external terminal adapter may be used as well. The terminal adapter 10A provides a protocol (physical layer and intermediate layer) conversion function for converting signal protocols such as V.35, RS-232, Universal Serial Bus (USB), IEEE 1394 (FireWire), etc. to an ISDN compliant protocol over a 4-wire interface. In the source facility 1, the bonding or multilink PPP mechanism may be incorporated in the source terminal 10, adapter 10A or in the NT1 14.
The NT1 14 connects the source facilities 1, via a two-wire line 15, to a switching module 26 located at the ISDN switch 22. Alternatively, a second network termination (NT2) may be used at the source facility 1 between NT1 and the terminal adapter to provide a switching and concentration function, such as with a digital private branch exchange (PBX). Likewise, the NT1 may be replaced with a NT12 that performs the functions of both the NT1 and NT2.
At the ISDN switch 22, the switching module 26 connects to a processor 24 and another switch module 28 via a bus 27, which allows digital commands and data to be passed between the respective switching modules 26 and 28, and the processor 24.
The equipment at the destination facility 2 may or may not be exactly similar to that of the source facilities 1. In the system shown at FIG. 1, the destination facility 2 includes the destination facsimile 16 having a terminal adapter 16A incorporated therein, which connects to another NT1 20 as shown. The NT1 20 connects to the switching module 28 in the ISDN switch 22, via another two-wire line 17 as shown.
ISDN communications is based on a seven layer protocol stack, as explained in reference to FIG. A.5 of Stallings, for example. Control signaling is accomplished between the respective user-network interface and occurs at a third layer of the protocol stack (i.e., the "network" layer) and is named I.451/Q.931. Thus, establishing and maintaining control signaling for a communication link established between the source facility 1 and a destination ISDN facility 2 through the D channel, and in particular, the ISDN network layer, data link layer and physical layer.
FIG. 2 is a frame structure 200 of a transmission from source facilities 1 to the ISDN switch 22, for an ISDN basic rate access. The frame structure 200 includes 48 bits which are transmitted in 250 .mu.sec. Components of the frame structure 200 include framing bits, F, dc balancing bits, L, B channel bits for the first B channel (16 per frame), B1, B channel bits for the second B channel (16 bits per frame), B2, D channel bits (4 per frame), D, auxiliary framing bit, Fa. A more detailed description of the frame structure, as well as a corresponding frame structure for the frames sent from the ISDN switch 22 to the source facilities 1, is described in Stallings, pp 212-215.
A link access protocol (LAPD) D channel is defined for establishing particular LAPD frames that are exchanged between the subscriber equipment (either at the source facility 1 or at the destination facility 2) and the ISDN switch 22. The call control protocol I.451/Q.931 is used on the D channel to establish, maintain and terminate connections on B channels.
FIG. 3 illustrates the signaling sequence between the source facility 1 and the ISDN switch 22. In order to establish each B channel connection between the source facility 1 and the destination facility 2, an initial communication link must be established on the D channel between the source facility 1 and the destination facility 2. To this end, a series of messages is sent back and forth between the source facilities 1 and the ISDN switch 22. This communication between the source facilities 1 and ISDN switch 22 occurs on a continuing basis on the D channel, while communications are maintained between the source facilities 1 and destination facilities 2 on the B channel. As shown in FIG. 3, several different messages are sent between the source facilities 1 and ISDN switch 22 while the D channel is maintained. A similar, redundant procedure is performed when the second B channel is established for bonding or multilink PPP purposes.
The direction of the arrows in FIG. 3 indicates a direction of communication between the source facilities 1 and the ISDN switch 22. The process for establishing a connection is initiated by the source facilities 1 by first sending a setup message. Particular features of the setup message will be discussed with respect to FIG. 3, however the purpose of the setup message is to provide general information regarding the request to connect to the ISDN switch 22. Next, the ISDN switch 22 responds with a call proceeding message that indicates that call establishment has been initiated. Subsequently, the ISDN switch 22, sends a connect message that indicates call acceptance by the source facilities 1. The source facilities 1 then sends a connect acknowledge signal that indicates the user has been awarded the call. When the user wishes to disconnect a call, the user sends a disconnect message via the source facilities 1 to the ISDN switch 22, requesting connection clearing. In response, a release message is sent from the ISDN switch 22, indicating the intent to release the channel and call reference. In response the source facilities 1 issues a release complete message, indicating that the release of the channel and the call reference. Subsequently, the call and information flow through the B channel is terminated.
FIG. 4 show the structure of a conventional ISDN D channel setup message. The setup message includes respective LAPD frames (e.g., 501, 503 . . . ) of different sizes (measured in octets). The message includes a flag frame 501 that is one octet in length, followed by a service access point identifier (SAPI) frame 503 having a command/response bit (CR) and address field extension bit (0). The SAPI frame 503 is joined with the terminal end point identifier (TEI) frame 505, each of which are one octet in length. A control frame 507, is one or two octets in length, and is followed by an information frame 509, which has a variable length between 0 and 128 octets. A frame check sequence frame 511 follows and occupies two octets in length. An end frame 513 serves as an end of setup message flag.
The SAPI frame 503 includes a first subfield "SAPI", that identifies a protocol layer-3 user, as well as subframes C/R and 0, that are used as a predetermined formatting feature of SAPI. The terminal end point identifier frame 505, is used to provide a unique terminal end point identifier that is used to identify the user's equipment. The control frame 507 defines the type of frame format that will be employed such as an information frame, supervisory frame, and unnumbered frame for example. The information frame 509, includes a variable number of octets varying from 0 to 128 and contains respective subfields that contain any sequence of bits that form an integral number of octets.
Thus, when a user wishes to send data to a destination, information in the information field is passed directly to the destination user without the ISDN switch deciphering the contents of the information. Following the information field 509, the frame check sequence 511 is included and forms an error-detection function by calculating a code from the remaining bits of the frame, exclusive of the flags. The normal code is a cyclical redundancy check code. Finally, the end flag frame 513, includes a specific code indicating the end of the setup message.
As identified by the present inventor, a limitation with the conventional ISDN setup architecture is that there is no suitable approach for arranging a single 128 kbps connection between a source terminal and a destination terminal, by way of the ISDN switch. Because the ISDN switch handles the different B channels independently, the ISDN switch imparts a significant degree of uncertainty regarding the communications paths assigned to different B channels that both have common origins and destinations--the net result being different, and perhaps non-static, interchannel delay. Conventional bonding and multilink PPP based systems overcome the delay obstacle imposed by the ISDN switch by employing more expensive and complex source and destination equipment so as to accommodate the processing and management overhead for "combining" two B channels. Furthermore, many conventional ISDN terminals such as G4 facsimile machines, are not configured to communicate over a 128 kbps link, as it is presumed that no more than 64 kbps is available for facsimile transmissions.